STEP BY STEP
The workflow
1
Convert your receipt photos to a single PDF
Drop in JPGs, HEICs, or PNGs from your phone. Each photo becomes one PDF page, auto-rotated, fitted to A4 or Letter — your choice. Add 0.25-inch margins so receipts don't bleed off the edge.
2
OCR so the totals are searchable
Phone-camera receipts are image-only. AI OCR adds a text layer that finance portals can index. Totals, vendor names, and dates become searchable.
3
Number the pages and add a cover
A cover page with your name, expense report number, and submission date turns 'IMG_4023.pdf' into a deliverable. Page numbers help finance reference specific receipts in queries.
WHY IT MATTERS
The case for doing this
Most expense systems accept one PDF per claim, not 27 separate JPGs. The 'attach 27 files' approach gets rejected by half of corporate portals and slows down the rest. One ordered, OCR'd PDF goes through in one upload, which means your reimbursement clears faster.
COMMON MISTAKES
What to avoid
Photographing receipts at angles
Phone photos taken at 30° angles don't OCR well — the text is distorted. Hold the phone flat over the receipt, parallel to it. Use a document-scanner app if you can.
Skipping the cover page
Finance teams reject anonymous PDFs. Always include a cover with your name, employee ID, and expense-report number.
Ordering chronologically when the trip is
Some finance systems want receipts ordered by date. Others want them grouped by category (transport, meals, lodging). Check before you compile.
TIPS FOR THE BEST RESULTS
What works
Take photos against a dark surface
White receipts on white tables don't auto-crop well. A dark surface gives the OCR engine clear edges.
Photograph the back of the receipt only if needed
Most receipts are one-sided. Don't double the page count for nothing.
Bundle by category if your portal supports it
Concur and Expensify let you upload one PDF per expense category. That's faster than one giant report.
RELATED WORKFLOWS
Other jobs people do with pdfcraft ai
FAQ
Questions about this workflow
Yes. The JPG-to-PDF conversion runs in your browser. OCR uploads but deletes within 60 minutes and never trains models.
Try the workflow on a real PDF
Free tools run in your browser. No signup, no watermarks.
Start with step 1